Lenny's Polls
How AI changes hiring
What 150+ product leaders have to say
The tl;dr
- Hiring leaders have no clear policy on AI. 43% of hiring leaders have no clear policy on candidate AI use. 26% have rebuilt interviews around evaluating AI workflows.
- Output is up, but headcount is flat. 56% report either same team with more output, or smaller team with same output. AI shows up as a productivity multiplier, not a hiring accelerant.
- Despite hiring process re-designs given AI shifts, time from posting to offer hasn't changed.
- Telling the human from the AI and catching real-time Zoom assistance dominate the pain points. Resumes have flattened into sameness. Teams respond by probing workflows and asking for shipped examples.
of teams still have no clear policy on AI in interviews
43%
Just 7% prohibit AI outright. Another 12% provide tools and evaluate how candidates use them.
How AI is reshaping hiring
What's your policy on candidates using AI during interviews?
Most teams haven't picked a stance on candidate AI use. 42.7% admit no clear policy. 12.2% provide tools and evaluate usage. 6.9% prohibit AI outright. The remaining 38% sit in the middle — allowing AI without scoring it, or redesigning exercises. Candidates show up not knowing what's allowed, what's scored, or what counts as cheating. The fastest move isn't a new rubric. It's a one-paragraph policy in the recruiting email before the next loop opens.
What's the single biggest change you've made to your interview process because of AI?
Hiring practices haven't caught up to the tools. Only 25.8% evaluate candidates' AI workflows. 18.9% redesigned take-homes to assume AI use. 12.9% added a live AI exercise. 18.2% haven't changed anything. Candidates arrive with AI fluency that didn't exist 18 months ago, but most interview loops still look pre-AI. Pick one stage to redesign before the next req opens. Don't wait for a stable rubric that may never arrive.
How has AI affected your team's headcount relative to output in the past year?
One in five teams is leaner. Most are doing more with the same people. 21.3% have shrunk teams while keeping or growing output. 34.6% kept the same team and produce more. 19.7% say AI hasn't moved the needle. 15.0% grew the team and credit AI. The "leaner" story is real for a fifth of teams; for most others, AI shows up as more output per person, not fewer people. Before approving the next backfill, ask whether AI plus the current team can absorb the work.
What's the biggest hiring mistake you've made (or seen) related to AI?
Mistakes split evenly between over-indexing and under-indexing on AI. Among teams who made or saw a hiring mistake, 15.2% over-weighted AI fluency at the cost of domain knowledge, 15.2% hired someone who talked a big AI game but couldn't deliver, and 9.6% regretted not testing for AI. AI talk is cheap. Shipped work with AI is the signal. Ask candidates for two things they shipped with AI last month, and what they'd do differently.
How has AI changed the seniority mix you're hiring for?
The senior tilt is real but not universal. 41.8% skew senior (20.9% almost exclusively, 20.9% mostly). 15.5% find AI-native juniors more valuable than before. 38.0% report no change. Two strategies work: hire seniors who use AI to compress execution, or hire AI-native juniors who ramp faster than old curves predict. The strategy that doesn't work is hiring juniors and assuming they'll pick up AI on the job.
Has AI changed the speed of your hiring process — from posting to signed offer?
AI tools haven't made hiring faster yet. 69.8% report no change in time from posting to signed offer. 11.9% say hiring has slowed; 18.3% report any speedup. The same teams reporting large output gains in engineering and product aren't seeing them in their hiring funnel. Added AI evaluation steps, screening against AI-assisted candidates, and the difficulty of telling humans from models are absorbing whatever speed AI screening tools could deliver.
How teams are evaluating AI fluency vs. the hardest parts of hiring in the ai era
Themes from open-ended responses. Click any to see quotes.
How teams are evaluating AI fluency
1
Probe real workflows and tool stacks
17
Show quotesHide quotes
"Diagnose their workflows, thinking, tech stack, and do a take home 48hr prototype that they are encouraged to build with AI"
"Seeing what tools they've integrated into their workflow"
"Ask them directly how they use AI in their daily lives and in their job"
2
Ask for specific shipped examples
12
Show quotesHide quotes
"“Walk me through a recent example of how you used AI to deliver a product. What was your process from start to finish?”"
"Tell me (or better show me) a specific project you worked on and use AI. What role did AI play, and what was your role? How was it received by customers? How did you evaluate how well it worked? Be specific."
"Ask about something they have built with AI that they are proud of."
3
Live build or take-home with AI
10
Show quotesHide quotes
"Give then a throwaway API key from Claude and ask them to deliver a real solution, fully-fledged, with just the business case. We are looking for people who: ship something that is desirable and viable; explore the solution space thoroughly before deciding on a course of action; and, lastly, document their choices and use proper information architecture that would compound."
"Take homes have unlimited AI help, we expect to see the result of how they use the tools, not which tools they are using and why."
"Watch them using the tools in live exercise. Ask them how they use tools in their workflow and how they find and try new tools."
4
Judgment over output
7
Show quotesHide quotes
"Theres direct and indirect signals... you can see how much of their thinking they offload to AI vs use it as something that augments their skills."
"We ask for specific examples of how people have used AI and dig into details. In particular, we are looking at how they evaluate AI output (both in product features and in their own work)"
"Behavioral interview questions to assess their usage of AI, probing for opinions, learnings, failures. I am listening for a new tool or approach. It's still about demonstrating impact, learning, curiosity and value."
The hardest parts of hiring in the AI era
1
Telling the human from the AI
13
Show quotesHide quotes
"how much is original / natural thinking vs. AI-aided"
"It's really hard to tell sometimes what is the person and what is AI. All inconsistencies seem suspicious, the water feels so muddled it's relatively hard to see any truth."
"Getting proof and validation that their work is genuinely theirs and not something the AI held their hand through."
2
Real-time cheating on Zoom
8
Show quotesHide quotes
"Live conversations on zoom when remote. Find that 50% of contingent roles are using interviewing aides that are giving them answers in real time."
"Many of them are actively using AI live during the interview, and then answering the questions I ask directly using AI, making it more difficult to screen and see if they understand the terms or actually have the experience we need."
"Their responses were canned and robotic, and it didn't inspire confidence that they can have real conversations with real people, which is a massive requirement of the job"
3
Resumes and applications all look the same
9
Show quotesHide quotes
"All CVs look the same."
"Weeding through AI crafted resumes."
"Too many top-of-funnel applications, hard to filter who to interview"
4
AI fluency is a moving target
7
Show quotesHide quotes
"The technology is developing so fast that it makes developing a meaningful rubric for AI aptitude difficult to construct."
"Nobody knows AI but everybody says they do and the field and tools are changing every day. It's really hard to target a specific skillset with regards to AI so I think we're going to put a lot more emphasis on problem solving."
"Getting through the hype. Everyone looks great on paper but they can't all keep it up when you get into details."